Use of personal care products during pregnancy and birth outcomes - A pilot study

Abstract

Prenatal exposure to endocrine disrupting chemicals (EDCs) from personal care products may be associated with birth outcomes including preterm birth and low birth weight. There is limited research examining the role of personal care product use during pregnancy on birth outcomes. Our pilot study consisted of 164 participants in the Environmental Reproductive and Glucose Outcomes (ERGO) study (Boston, MA), with data on self-reported personal care product use at four study visits throughout pregnancy (product use in the 48 h before a study visit and hair product use in the month before a study visit). We used covariate-adjusted linear regression models to estimate differences in mean gestational age at delivery, birth length, and sex-specific birth weight-for-gestational age (BW-for-GA) Z-score based on personal care product use. Hair product use in the past month prior to certain study visits was associated with decreased mean sex-specific BW-for-GA Z-scores. Notably, hair oil use in the month prior to study visit 1 was associated with a lower mean BW-for-GA Z-score (V1: -0.71, 95% confidence interval: -1.12, -0.29) compared to non-use. Across all study visits (V1-V4), increased mean birth length was observed among nail polish users vs. non-users. In comparison, decreased mean birth length was observed among shave cream users vs. non-users. Liquid soap, shampoo, and conditioner use at certain study visits were significantly associated with higher mean birth length. Suggestive associations were observed across study visits for other products including hair gel/spray with BW-for-GA Z-score and liquid/bar soap with gestational age. Overall, use of a variety of personal care products throughout pregnancy was observed to be associated with our birth outcomes of interest, notably hair oil use during early pregnancy. These findings may help inform future interventions/clinical recommendations to reduce exposures linked to adverse pregnancy outcomes.

Keywords: Birth outcomes; Endocrine disrupting chemicals; Hair products; Personal care products; Pregnancy.

Figure 1.

Self-reported use of (A) personal care products within the past 48 hours and (B) hair products within the past month prior to each pregnancy study visit (V1-4) and “ever” in pregnancy (n=164*). “Ever” represents self-reported use at any of the four pregnancy study visits. *Sample size varied across visits and by product category due to missing data (range: n=137 to n=164).

Figure 2.

Self-reported frequency of hair product use (daily,

Figure 3.

Covariate-adjusted differences in mean birth outcomes associated with use of personal care products compared to non-use within the 48 hours prior to each pregnancy visit (n=164). Figures depict the estimated mean difference in (A) gestational age at delivery (days), (B) birth length (cm), and (C) sex-specific BW-for-GA Z-score in product users compared to non-users at each study visit and “ever” users compared to “never” users during pregnancy. Models were adjusted for the following covariates: (A) parental age at consent (years), bachelor’s degree or higher (yes vs. no), BMI at visit 1 (kg/m²), and parity (0, 1, 2+); (B) parental age at consent (years), bachelor’s degree or higher (yes vs. no), BMI at visit 1 (kg/m²), parity (0, 1, 2+), infant sex, and gestational age at delivery (weeks); (C) parental age at consent (years), bachelor’s degree or higher (yes vs. no), BMI at visit 1 (kg/m²), and parity (0, 1, 2+). N=3 participants missing birth length data.

Figure 4.

Covariate-adjusted differences in mean sex-specific BW-for-GA Z-score associated with (A) hair product use and (B) frequency of hair product use within the month prior to each pregnancy visit (n=149*). All models were adjusted for parental participant age at consent (years), bachelor’s degree or higher (yes vs. no), BMI at visit 1 (kg/m²), and parity (0, 1, 2+). (B) “Ever” use categories represent report of daily use or less than daily use at one or more pregnancy visits compared to “never” users. Visit-specific estimates for frequency of hair lotion and leave-in conditioner use were not modeled due to insufficient numbers of daily users at individual study visits. *Due to missing frequency data at some individual visits sample sizes ranged from n=137 to n=149.

Figure 5.

Estimated normal probability density functions of sex-specific BW-for-GA Z-scores in daily, less than daily, and non-users of hair oil at Visit 1 in pregnancy (n=144). Vertical lines depict the mean BW-for-GA Z-scores for daily users (mean ± SD: −1.2 ± 1.5), less than daily users (−0.7 ± 0.9) non-users (−0.09 ± 1.1).